Robots that can adapt like animals
Open accessThe DOI matches its Crossref record; 39 of the 40 references it cites were found in a registry or an index.
Abstract
Robots have transformed many industries, most notably manufacturing, and have the power to deliver tremendous benefits to society, such as in search and rescue, disaster response, health care and transportation. They are also invaluable tools for scientific exploration in environments inaccessible to humans, from distant planets to deep oceans. A major obstacle to their widespread adoption in more complex environments outside factories is their fragility. Whereas animals can quickly adapt to injuries, current robots cannot 'think outside the box' to find a compensatory behaviour when they are damaged: they are limited to their pre-specified self-sensing abilities, can diagnose only anticipated failure modes, and require a pre-programmed contingency plan for every type of potential damage, an impracticality for complex robots. A promising approach to reducing robot fragility involves having robots learn appropriate behaviours in response to damage, but current techniques are slow even with small, constrained search spaces. Here we introduce an intelligent trial-and-error algorithm that allows robots to adapt to damage in less than two minutes in large search spaces without requiring self-diagnosis or pre-specified contingency plans. Before the robot is deployed, it uses a novel technique to create a detailed map of the space of high-performing behaviours. This map represents the robot's prior knowledge about what behaviours it can perform and their value. When the robot is damaged, it uses this prior knowledge to guide a trial-and-error learning algorithm that conducts intelligent experiments to rapidly discover a behaviour that compensates for the damage. Experiments reveal successful adaptations for a legged robot injured in five different ways, including damaged, broken, and missing legs, and for a robotic arm with joints broken in 14 different ways. This new algorithm will enable more robust, effective, autonomous robots, and may shed light on the principles that animals use to adapt to injury.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source doesn’t support the claim as written.
In the citing paper
“We render images using the MuJoCo Physics Engine's [45] built-in renderer. This renderer is not intended to be photo-realistic, and physically plausible choices of textures and lighting are not needed.”
From Domain randomization for transferring deep neural networks from simulation to the real world (2017) · cites this paper as [45]
Checked by a reasoning model — may be wrong.
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“under Quality-Diversity optimization (Pugh et al., 2016), the search space is divided into behavioral niches to evolve diverse solutions, a paradigm that challenged the dominant approach of training a single agent, as it proved more robust in real-world settings (Cully et al., 2015).”
From Dynamics of niche construction in adaptable populations evolving in diverse environments (2023) · cites this paper as Cully et al., 2015
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Similarly, extensive reliance on simulation and offline training further distances learning from the full variability and constraint structure of real sensorimotor interaction (Jakobi et al., 1995; Cully et al., 2015).”
From Toward Enactive Artificial Intelligence (2026) · cites this paper as Cully et al., 2015
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Gaussian Processes for Machine Learning
Rasmussen, C. E., & Williams, C. K. I. (2005) · doi:10.7551/mitpress/3206.001.0001 · Cited by 11,705
Corke, P., Roberts, J., Cunningham, J., & Hainsworth, D. (2008) · Springer Handbook of Robotics · doi:10.1007/978-3-540-30301-5_50 · Cited by 3,543
A survey of robot learning from demonstration
Argall, B. D., Chernova, S., Veloso, M., & Browning, B. (2009) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2008.10.024 · Cited by 3,227
Reinforcement learning in robotics: A survey
Kober, J., Bagnell, J. A., & Peters, J. (2013) · The International Journal of Robotics Research · doi:10.1177/0278364913495721 · Cited by 3,160 · cited 9×
“The key insight is that, whereas current learning algorithms either start with no knowledge of the search space 17 or with minimal knowledge from a few human demonstrations 17,18 , animals better understand the space of possible behaviors and their value from previous experience 19 , enabling…”
BackgroundMethodResultInfluentialDiagnosis and Fault-Tolerant Control
Blanke, M., Kinnaert, M., Lunze, J., & Staroswiecki, M. (2003) · Springer eBooks · doi:10.1007/978-3-662-05344-7 · Cited by 2,789
Nonlinear Programming: Sequential Unconstrained Minimization Techniques
Fiacco, A. V., & McCormick, G. P. (1968) · Defense Technical Information Center (DTIC) · Cited by 2,262
Stanley: The robot that won the DARPA Grand Challenge
Thrun, S., Montemerlo, M., Dahlkamp, H., Stavens, D., Aron, A., Diebel, J., Fong, P., Gale, J., Halpenny, M., Hoffmann, G., et al. (2006) · Journal of Field Robotics · doi:10.1002/rob.20147 · Cited by 2,237 · cited 2×
“3C, undam-aged: 24 [16; 41] seconds, 3 [2; 5] physical trials, repaired: 29 [16; 82] seconds, 3.5 [2; 10] trials).”
BackgroundInfluentialBrochu, E., Cora, V. M., & de Freitas, N. (2010) · arXiv (Cornell University) · doi:10.48550/arxiv.1012.2599 · Cited by 2,011
Bayesian integration in sensorimotor learning
Körding, K. P., & Wolpert, D. M. (2004) · Nature · doi:10.1038/nature02169 · Cited by 2,004 · cited 1×
“…optimization procedure followed by our robot appears similar to the technique employed by humans when they optimize an unknown function 23 , and there is strong evidence that animal brains learn probability distributions, combine them with prior knowledge, and act as Bayesian optimizers 25,26 .”
MethodControl of mental activities by internal models in the cerebellum
Ito, M. (2008) · Nature Reviews Neuroscience · doi:10.1038/nrn2332 · Cited by 1,356 · cited 1×
“A final parallel is that the simulator and Gaussian process components of Intelligent Trial and Error are two forms of predictive models, which are known to exist in animals 29,12 .”
Research works printed but not linked: 2
- Siciliano, B. & Khatib, O. Springer Handbook of Robotics (Springer, 2008). · doi:10.1007/978-3-540-30301-5
- Kluger, J. & Lovell, J. Apollo 13 (Mariner Books, 2006).
Cited by
Items 501–600 of 1,007
Cazenille, L., Bredèche, N., & Halloy, J. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-24741-6_4 · Cited by 11
Fast Adaptation to New Environments via Policy-Dynamics Value Functions
Răileanu, R., Goldstein, M. A., & Szlam, A. D. (2020) · International Conference on Machine Learning · Cited by 31
Embracing curiosity eliminates the exploration-exploitation dilemmaPreprint
Peterson, E. J., & Verstynen, T. D. (2019) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/671362 · Cited by 15
High-Dimensional Bayesian Optimization via Nested Riemannian ManifoldsPreprint
Jaquier, N., & Rozo, L. D. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2010.10904 · Cited by 35
“Introducing domain knowledge into surrogate models and acquisition functions has recently shown to improve the performance and scalability of BO [11, 3, 45, 30, 14].”
BackgroundGavra, V., & Kampen, E.-J. V. (2024) · Journal of Guidance Control and Dynamics · doi:10.2514/1.g008112 · Cited by 9
Du, Q., Wang, H., Jiang, B., & Wang, X. (2025) · Briefings in Bioinformatics · doi:10.1093/bib/bbaf286 · Cited by 13
“…conditions [33, 34], virtual screening of drugs against specific targets [35, 36], optimization of fast-charging protocols for batteries [37], optimization of tasks for dissociating frontoparietal brain networks [38], and the design of gaits for legged robots to adapt to injuries [39], etc.”
MAP-Elites with Cosine-Similarity for Evolutionary Ensemble Learning
Zhang, H., Chen, Q., Tonda, A. P., Xue, B., Banzhaf, W., & Zhang, M. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-29573-7_6 · Cited by 9
“In recent years, quality diversity (QD) optimization has been widely used to tackle the problem of deceptive landscapes [36] and produce diverse solutions [9].”
BackgroundMany-task evolutionary algorithm with adaptive knowledge transfer via density-based clustering
Bao, C., Gao, D., Ding, Y., Xu, L., & Goodman, E. D. (2023) · Knowledge-Based Systems · doi:10.1016/j.knosys.2023.110906 · Cited by 9
You, B., Chen, X., Li, J., Ding, L., & Dong, Z. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.112551 · Cited by 9
Learning Based MPC for Autonomous Driving Using a Low Dimensional Residual Model
Li, Y., Huang, C. H., Yang, D., Liu, W., & Li, J. (2025) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra55743.2025.11128642 · Cited by 9
“In recent years, data-driven algorithms have received significant attention in the academic community due to their excellent ability to approximate unknown functions [12], [13].”
Evolving Robots on Easy Mode: Towards a Variable Complexity Controller for Quadrupeds
Nygaard, T. F., Martin, C. P., Tørresen, J., & Glette, K. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-16692-2_41 · Cited by 9
Ratté, S. (2019) · Environment and Society · doi:10.3167/ares.2019.100109 · Cited by 9
Balachandar, P., & Michmizos, K. P. (2020) · IEEE/RAS-EMBS International Conference on Biomedical Robotics and Biomechatronics (BioRob) · doi:10.1109/biorob49111.2020.9224303 · Cited by 10
“new controllers capable of naturally handling richer, noisier and more complex scenarios [44].”
BackgroundExploring Robot Personality through Big Data Mining: A Century-Long Analysis from Google Books
Xu, L., & Chao, C. (2023) · International Journal of Human-Computer Interaction · doi:10.1080/10447318.2023.2266796 · Cited by 9
“This shift in focus from mere “intellectual” capabilities to “intelligent” capabilities reflected the aspiration of creating machines that not only processed information but also had the capacity to learn, reason, and adapt to novel situations (Cully et al., 2015).”
Hybridizing novelty search for transfer learning
Didi, S., & Nitschke, G. (2016) · IEEE Symposium Series on Computational Intelligence (SSCI) · doi:10.1109/ssci.2016.7850180 · Cited by 9
“In Evolutionary Robotics (ER) [1] there is increasing empirical evidence that maintaining diversity in genotypes (controller encodings) and phenotypes (controller behaviors) improves the quality (task performance) of evolved behaviors [2], [3], [4], [5].”
BackgroundDeep Reinforcement Learning for Single-Shot Diagnosis and Adaptation in Damaged Robots
Verma, S., Nair, H., Agarwal, G., Dhar, J., & Shukla, A. (2020) · Joint International Conference on Data Science & Management of Data (CODS-COMAD) · doi:10.1145/3371158.3371168 · Cited by 11
“Rather than considering two separate phases for damage diagnosis and recovery algorithm generation, Cully et al. [Cully et al. 2015], proposed a method inspired from animals, who perform trial and error to determine the least painful alternate gait in the presence of injury.”
BackgroundMethodInfluentialSelf-engineering – Technological Challenges
Roy, R., & Brooks, S. (2020) · Lecture notes in networks and systems · doi:10.1007/978-3-030-46817-0_2 · Cited by 9
“A similar approach with a six-legged robot was tested more recently with optimised algorithms, reducing the time needed to adapt to minutes [29].”
BackgroundMethodDiscovering and Exploiting Sparse Rewards in a Learned Behavior Space
Paolo, G., Coninx, A., Laflaquière, A., & Doncieux, S. (2024) · Evolutionary Computation · doi:10.1162/evco_a_00343 · Cited by 12
“This gave rise to a new family of methods called quality-diversity (QD) (Cully and Demiris, 2017; Pugh et al., 2016; Cully et al., 2015; Eysenbach et al., 2018; Lehman and Stanley, 2011; Paolo et al., 2021; Mouret and Clune, 2015).”
Severe damage recovery in evolving soft robots through differentiable programming
Horibe, K., Walker, K., Palm, R. B., Sudhakaran, S., & Risi, S. (2022) · Genetic Programming and Evolvable Machines · doi:10.1007/s10710-022-09433-z · Cited by 13
Self Exploration of Sensorimotor Spaces in Robots.
Benureau, F. (2015) · theses.fr (ABES) · doi:10.70675/1d482feaz6e01z4426zbc1cz25388edcb189 · Cited by 19
The Role of Stepping Stones in MAP-Elites: Insights from Search Trajectory Networks
Nadizar, G., Rusin, F., Medvet, E., & Suazo, G. S. O. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-89991-1_14 · Cited by 10
Transferable Trial-Minimizing Progressive Peg-in-hole Model
Ding, J., Wang, C., & Lu, C. (2019) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros40897.2019.8968282 · Cited by 8
“to humans, robots can also actively improve their estimation on a certain problem through a series of trials [18].”
BackgroundHu, S., Zhang, W., Li, D., & Wu, B. (2023) · Physica A Statistical Mechanics and its Applications · doi:10.1016/j.physa.2023.128810 · Cited by 8
Towards Multi-Morphology Controllers with Diversity and Knowledge Distillation
Mertan, A., & Cheney, N. (2024) · Genetic and Evolutionary Computation Conference (GECCO) · doi:10.1145/3638529.3654013 · Cited by 11
“As opposed to methods like [2, 8, 19] where the system spends time to figure out how to recover from damage, we would be measuring instant (i.e. zero-shot) damage recovery through the robustness of the controller.”
MethodZehouani, N., Ababou, M., & Faquir, S. (2024) · Lecture notes in networks and systems · doi:10.1007/978-3-031-68650-4_39 · Cited by 9
A review on corrosion modelling for submarine pipeline
Zhao, Z., Akhtar, M. N., Bakar, E. A., & Razak, N. B. A. (2025) · Ain Shams Engineering Journal · doi:10.1016/j.asej.2025.103411 · Cited by 10
Abstraction as a Mechanism to Cross the Reality Gap in Evolutionary Robotics
Scheper, K. Y. W., & de Croon, G. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-43488-9_25 · Cited by 10
“Some resent work has shown promising results in bridging the reality gap using abstracted methods [5,7].”
BackgroundAn Analysis of Phenotypic Diversity in Multi-solution Optimization
Hagg, A., Preuß, M., Asteroth, A., & Bäck, T. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-63710-1_4 · Cited by 9
“QD, reintroducing objectives [3,12], finds a diverse set of high quality optimizers by performing niching in phenotypic space.”
BackgroundFault Tolerant Control in Shape-Changing Internal Robots
Balasubramanian, L., Wray, T., & Damian, D. D. (2020) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra40945.2020.9196989 · Cited by 10
MAP-Elites for Constrained Optimization
Fioravanzo, S., & Iacca, G. (2021) · Springer eBooks · doi:10.1007/978-981-33-6710-4_7 · Cited by 11
Modeling, Simulation and Implementation of All Terrain Adaptive Five DOF Robot
Wang, Z., Zhao, J., & Zeng, G. (2022) · Sensors · doi:10.3390/s22186991 · Cited by 8
“developed an intelligent trial-and-error algorithm that enables a multipedal robot to adapt to impairments in its environment and it takes less than two minutes in a large search space without self-diagnosis or contingency planning, demonstrating robust adaptation when some of the multipedal robot’s legs are damaged p…”
BackgroundAutomatic Generation of Robot Facial Expressions with Preferences
Tang, B., Cao, R., Chen, R., Chen, X., Hua, B., & Wu, F. (2023) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48891.2023.10160409 · Cited by 11
“Based on this, we proposed a general learning framework that utilizes MAP-Elites [58] for automatically generating specified facial expressions.”
BackgroundMethodInfluentialEmbodied essentialism in the reconstruction of the animal sign in robot animal design
Torres–Martínez, S. (2024) · Biosystems · doi:10.1016/j.biosystems.2024.105178 · Cited by 8
Evolving robot controllers for a bio-hybrid system
Mariano, P., Salem, Z., Mills, R., Schönwetter-Fuchs-Schistek, S., Correia, L. M. P., & Schmickl, T. (2018) · Conference on Artificial Life (ALIFE) · doi:10.1162/isal_a_00036 · Cited by 8
“Conventional ER has diverse applications including controllers of robots that resemble or mimic animals (e.g., Cully et al., 2015).”
BackgroundCreativity: a Driver for Research on Robotics in Open Environments
Doncieux, S. (2016) · Intellectica Revue de l Association pour la Recherche Cognitive · doi:10.3406/intel.2016.1795 · Cited by 9
Towards an intelligent photonic system
Zou, W., Ma, B., Xu, S., Zou, X., & Wang, X. (2020) · Science China Information Sciences · doi:10.1007/s11432-020-2863-y · Cited by 8
Behavioral Diversity Generated From Body–Environment Interactions in a Simulated Tensegrity Robot
Terajima, R., Inoue, K., Yonekura, S., Nakajima, K., & Kuniyoshi, Y. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2021.3139083 · Cited by 8
Gradient-Informed Quality Diversity for the Illumination of Discrete Spaces
Boige, R., Richard, G., Donà, J., Pierrot, T., & Cully, A. (2023) · Genetic and Evolutionary Computation Conference (GECCO) · doi:10.1145/3583131.3590407 · Cited by 11
“This core idea showed great outcomes in different fields such as robotics [8] where it allows to learn to control robots using diverse and efficient policies or latent space exploration [15] of generative models to generate a diverse set of high quality images or video game levels.”
BackgroundSelf-citationModeling user selection in quality diversity
Hagg, A., Asteroth, A., & Bäck, T. (2019) · Genetic and Evolutionary Computation Conference (GECCO) · doi:10.1145/3321707.3321823 · Cited by 12
“ee a solution is functional. Novelty Search with Local Competition (NSLC) [12] firstly combined quality and diversity, and a few years later Multidimensional Archive of Phenotypic Elites (MAP-Elites) [5] was introduced. The main differences between the methods are the way they define an archive and how they select the…”
BackgroundMethodInfluentialHarnessing Distribution Ratio Estimators for Learning Agents with Quality and DiversityPreprint
Gangwani, T., Peng, J., & Zhou, Y. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2011.02614 · Cited by 12
“Having a repertoire of skills is useful when knowledege transfer is done to a target environment that has constraints on the set of feasible behaviors [10].”
BackgroundAn Energy-Saving Snake Locomotion Gait Policy Obtained Using Deep Reinforcement LearningPreprint
Liu, Y., & Farimani, A. B. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2103.04511 · Cited by 10
“prior knowledge of the environment [24].”
Overcoming Deceptiveness in Fitness Optimization with Unsupervised Quality-Diversity
Coiffard, L., Templier, P., & Cully, A. (2025) · Genetic and Evolutionary Computation Conference (GECCO) · doi:10.1145/3712256.3726314 · Cited by 9
Self-citationRobust and Adaptive Robot Self-Assembly Based on Vascular Morphogenesis
Soorati, M. D., Ghofrani, J., Zahadat, P., & Hamann, H. (2018) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2018.8594093 · Cited by 7
“Similar to studies of how individual robots [6] or networks of robots [7] adapt to damages and changes in their environment, we want to make a step towards adaptability in self-assembling robots.”
ResultAssessing the Effects of Interacting with MAP-Elites
Alvarez, A., Font, J., Dahlskog, S., & Togelius, J. (2021) · Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment · doi:10.1609/aiide.v17i1.18899 · Cited by 7
“MAP-Elites have shown excellent results at generating QD individuals in games (Fontaine et al. 2020; Alvarez et al. 2020), and offline adaptation based on its generated repertoire (Cully et al. 2015; Gonzalez-Duque et al. 2020).”
BackgroundMethodA model-free method to learn multiple skills in parallel on modular robots
van Diggelen, F., Cambier, N., Ferrante, E., & Eiben, A. E. (2024) · Nature Communications · doi:10.1038/s41467-024-50131-4 · Cited by 9
“This involves problems due to catastrophic forgetting (a problem that arises in lifelong/(deep) reinforcement learning 30,31 ), or the curse of dimensionality (e.g., MAP-elites, 32,33 ), and overall time constraints 11,29 .”
Estimation of the legs’ state of a mobile robot based on Long Short-Term Memory network
Albadin, A., Albitar, C., & Saba, M. A. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2024.109539 · Cited by 7
From single-objective to multi-objective reinforcement learning-based model transformation
Eisenberg, M., & Wimmer, M. (2024) · Software & Systems Modeling · doi:10.1007/s10270-024-01233-6 · Cited by 8
“If states and outcomes of actions can be simulated for a domain, however, this can be utilized for training, as demonstrated by some work on robots for RL tasks [46–49].”
Neurobiologically Inspired Self-Monitoring Systems
Chiba, A. A., & Krichmar, J. L. (2020) · Proceedings of the IEEE · doi:10.1109/jproc.2020.2979233 · Cited by 10
“On the one hand, this causes the system to adjust its actions based [11].”
BackgroundMethodBehavioral Repertoire via Generative Adversarial Policy Networks
Jegorova, M., Doncieux, S., & Hospedales, T. M. (2020) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2020.3008574 · Cited by 8
“Exploration within a set of diverse historical behaviors that solved a task can quickly lead to a solution that succeeds in a new environment [4].”
InfluentialIdeal Mechanization: Exploring the Machine Metaphor through Theory and Performance
LaViers, A. (2019) · Arts · doi:10.3390/arts8020067 · Cited by 8
A Polyester/Polypyrrole Textile‐Based Ultrasensitive Wearable Microdistance Sensor
Chen, R., Ma, X., Chai, Y., Hua, K., Gui, Q., He, Y., Yang, X., Zhu, N., Wang, Y., Ji, J., et al. (2021) · Macromolecular Materials and Engineering · doi:10.1002/mame.202100478 · Cited by 7
Surana, S., Lim, B., & Cully, A. (2023) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48891.2023.10161267 · Cited by 9
“monly been used to learn a diverse repertoire of primitive controllers [33] which can then be used effectively for downstream applications such a damage recovery [25], [34] or with planning algorithms to perform longer horizon tasks”
BackgroundMethodInfluentialSelf-citationWang, R.-J., Xue, K., Shang, H., Qian, C., Fu, H., & Fu, Q. (2023) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2023/482 · Cited by 15
“Quality-Diversity (QD) algorithms [Cully et al. , 2015; Mouret and Clune, 2015; Cully and Demiris, 2018; Chatzi-lygeroudis et al. , 2021] are a specific type of EAs that aim to return a set of high-quality and diverse solutions in a single run.”
BackgroundMethodInfluentialTowards Jumping Skill Learning by Target-guided Policy Optimization for Quadruped Robots
Zhang, C., Zou, W., Cheng, N., & Zhang, S. (2024) · Machine Intelligence Research · doi:10.1007/s11633-023-1429-5 · Cited by 8
Investigating the investment readiness of European SMEs: A machine learning approach
Alexakis, C. A., Gogas, P., Petrella, G., Polemis, M., & Salvadè, F. (2025) · International Review of Financial Analysis · doi:10.1016/j.irfa.2025.104439 · Cited by 7
Evolving morphological robustness for collective robotics
Putter, R., & Nitschke, G. (2017) · IEEE Symposium Series on Computational Intelligence (SSCI) · doi:10.1109/ssci.2017.8285267 · Cited by 9
Honeybees Prefer to Steer on a Smooth Wall With Tetrapod Gaits
Zhao, J., Zhu, F., & Yan, S. (2018) · Journal of Insect Science · doi:10.1093/jisesa/iey038 · Cited by 7
Open-ended evolution with multi-containers QD
Doncieux, S., & Coninx, A. (2018) · Genetic and Evolutionary Computation Conference Companion (GECCO Companion) · doi:10.1145/3205651.3205705 · Cited by 8
“They can be used in a two-steps process in which (1) a diverse set of solutions is generated offline before (2) the agent can exploit it online to find the behavior that fits to the situation [2, 4–6].”
BackgroundRakicevic, N., & Kormushev, P. S. (2019) · Autonomous Robots · doi:10.1007/s10514-019-09842-7 · Cited by 10
Schilling, M., Ritter, H. J., & Ohl, F. W. (2019) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc.2019.8914302 · Cited by 7
“[5] introduced a specific, low-dimensional behavioral space that could act as a representational structure for finding stepping stones.”
BackgroundInfluentialResilient Task Planning and Execution for Reactive Soft Robots
Hamill, S., Whitehead, J., Ferenz, P., Shepherd, R. F., & Kress‐Gazit, H. (2019) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2019.8794303 · Cited by 7
“Previous research has addressed the issue of gait resilience with respect to walking robots [14], [15] as well as soft robots [16], but these methods do not provide guarantees of robot performance.”
BackgroundMarjaninejad, A., Tan, J., & Valero‐Cuevas, F. J. (2020) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc44109.2020.9176089 · Cited by 8
“, contact dynamics, injury to the body, or changes in the environment) [25], [23], [34], [35], [36], [37].”
Automating Representation Discovery with MAP-Elites
Gaier, A., Asteroth, A., & Mouret, J.-B. (2020) · Cited by 7
Self-citationTarokh, M. (2020) · Autonomous Robots · doi:10.1007/s10514-020-09929-6 · Cited by 7
Scalable and Parallel Deep Bayesian Optimization on Attributed Graphs
Cui, J., Yang, B. X., Sun, B., Hu, X., & Liu, J. (2020) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2020.3027552 · Cited by 10
Optimizing metal grating back reflectors for III-V-on-silicon multijunction solar cells
Tillmann, P., Bläsi, B., Burger, S., Hammerschmidt, M., Höhn, O., Becker, C., & Jäger, K. (2021) · Optics Express · doi:10.1364/oe.426761 · Cited by 7
Data-driven predictive modeling of Hubble parameter
Saltı, M., Ciger, E., Kangal, E. E., & Zengin, B. (2022) · Physica Scripta · doi:10.1088/1402-4896/ac807c · Cited by 7
Assessing evolutionary terrain generation methods for curriculum reinforcement learning
Howard, D., Munn, H., Dolcetti, D., Kannemeyer, J., & Robinson, N. (2022) · Genetic and Evolutionary Computation Conference (GECCO) · doi:10.1145/3512290.3528870 · Cited by 9
“Also in an evolutionary context, quality-diversity algorithms [27] present an effective method for storing and traversing a curriculum, shown in both robotics and game-playing contexts [9] [4].”
MethodInformation Optimization and Transferable State Abstractions in Deep Reinforcement Learning
Gomez, D., Quijano, N., & Giraldo, L. F. (2022) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2022.3200726 · Cited by 12
“Instead of trying to remember everything [8], [71] or nothing, it could be more efficient remembering just the information that is common to most tasks and that facilitates discovering the optimal behavior.”
BackgroundImproving HybrID: How to best combine indirect and direct encoding in evolutionary algorithms
Helms, L., & Clune, J. (2017) · PLoS ONE · doi:10.1371/journal.pone.0174635 · Cited by 7
Self-citationLearning a Behavioral Repertoire from Demonstrations
Justesen, N., González-Duque, M., Cabarcas, D., Mouret, J.-B., & Risi, S. (2020) · IEEE Conference on Games (CoG) · doi:10.1109/cog47356.2020.9231897 · Cited by 7
“The Intelligent Trial-andError algorithm (IT&E) uses BO for robot adaptation to deal with changes in the environment by intelligently searching in the continuous behavioral space of policies found by MAPElites [31].”
BackgroundSelf-citationArtificial Immune Systems: An Overview for Faulting Actuators
Kidd, R. B. (2019) · Actuators · doi:10.3390/act8030053 · Cited by 7
The Elements of Flexibility for Task-Performing Systems
Mayer, S., Sotto, L. F. D. P., & Garcke, J. (2023) · IEEE Access · doi:10.1109/access.2023.3238872 · Cited by 8
“diversity in evolutionary algorithms [98], [99], [100], [101].”
BackgroundThe evolution and function of play
Schank, J. C. (2015) · Adaptive Behavior · doi:10.1177/1059712315611964 · Cited by 7
“Groos (1898), in the first book dedicated to animal play, argued that Spencer’s imitative theory could not explain why young animals play because they lack adult functional behaviors to imitate.”
BackgroundReal-World Robotic Perception and Control Using Synthetic Data
Tobin, J. (2019) · UC Berkeley · Cited by 16
Bibi, S., & Yang, L. (2025) · Discover Artificial Intelligence · doi:10.1007/s44163-025-00647-5 · Cited by 6
Bayesian optimisation for low-noise aerofoil design with aerodynamic constraints
Paruchuri, C. C., & Vellanki, P. (2020) · International Journal of Aeroacoustics · doi:10.1177/1475472x20978395 · Cited by 6
Using robots to investigate the evolution of adaptive behavior
Bongard, J. (2015) · Current Opinion in Behavioral Sciences · doi:10.1016/j.cobeha.2015.11.008 · Cited by 7
“In another project [46 ], a robot stored a large portfolio of diverse actions and successfully predicted which of those actions would be useful in the face of unanticipated events.”
BackgroundOn the Robustness and Reliability in the Pose Deformation System of Mobile Robots
Cui, J., Zeng, S., Ren, Y., Chen, X., & Gao, Z. (2018) · IEEE Access · doi:10.1109/access.2018.2835836 · Cited by 6
“[2] and their research team designed an intelligent self-adaption control algorithm for a six-legged robot.”
BackgroundThe Algorithmic Unconscious: How Psychoanalysis Helps in Understanding AI
Possati, L. M. (2021) · Cited by 7
A geometric semantic macro-crossover operator for evolutionary feature construction in regression
Zhang, H., Chen, Q., Xue, B., Banzhaf, W., & Zhang, M. (2023) · Genetic Programming and Evolvable Machines · doi:10.1007/s10710-023-09465-z · Cited by 6
“To achieve this goal, we propose using MAP-Elites, a niching algorithm that has shown promising results in many quality-diver-sity optimization tasks [37, 38].”
MethodWang, R.-J., Xue, K., Wang, Y., Yang, P., Fu, H., Fu, Q., & Qian, C. (2025) · Frontiers of Computer Science · doi:10.1007/s11704-025-41167-w · Cited by 8
“ME [39, 43], the most well-known QD algorithm, maintains an archive by discretizing the behavior space into M cells {S i } Mi = 1 and storing at most one solution in each cell.”
InfluentialBacharis, A., Polyzos, K. D., Nelson, H. J., Giannakis, G. B., & Papanikolopoulos, N. (2025) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2025.3625501 · Cited by 6
“BO has been adopted in a range of practical settings including hyperparameter tuning [14], drug discovery [15] and robotic-based tasks [16], [17], [18] to name a few.”
Deep dive into model-free reinforcement learning for underwater locomotion: theory and practice
Jiao, Y., Ling, F., Heydari, S., Heess, N., Merel, J., & Kanso, E. A. (2026) · Bioinspiration & Biomimetics · doi:10.1088/1748-3190/ae4930 · Cited by 6
Optimizing Morphology and Locomotion on a Corpus of Parametric Legged Robots
Passault, G., Rouxel, Q., Fabre, R., N’Guyen, S., & Ly, O. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-42417-0_21 · Cited by 6
Real-time robot tracking and following with neuromorphic vision sensor
Mishra, A. K., Ghosh, R., Goyal, A., Thakor, N. V., & Kukreja, S. L. (2016) · IEEE/RAS-EMBS International Conference on Biomedical Robotics and Biomechatronics (BioRob) · doi:10.1109/biorob.2016.7523451 · Cited by 6
“They also have the ability to rapidly adapt to different forms of injury [22].”
BackgroundRobots That Think Fast and Slow: An Example of Throwing the Ball Into the Basket
Li, T. S., Kuo, P., Ho, Y.-F., Liu, C.-Y., Yu, T.-C., Ye, Y.-T., Chang, C.-Y., Chen, G.-Y., Chien, C.-W., Chen, W., et al. (2016) · IEEE Access · doi:10.1109/access.2016.2601167 · Cited by 7
LaViers, A. (2019) · PLoS ONE · doi:10.1371/journal.pone.0215671 · Cited by 6
Bootstrapped Neuro-Simulation as a method of concurrent neuro-evolution and damage recovery
Leonard, B. A., du Plessis, M. C., & Woodford, G. W. (2019) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2019.103398 · Cited by 7
Automated, Explainable Rule Extraction from MAP-Elites Archives
Urquhart, N. B., Höhl, S., & Hart, E. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-72699-7_17 · Cited by 6
Safe trajectory optimization for whole-body motion of humanoids
Modugno, V., Nava, G., Pucci, D., Nori, F., Oriolo, G., & Ivaldi, S. (2017) · IEEE-RAS International Conference on Humanoid Robots (Humanoids) · doi:10.1109/humanoids.2017.8246958 · Cited by 6
“to test on the robot online by trial and error [31].”
BackgroundOnline Gait Adaptation of a Hexapod Robot Using an Improved Artificial Hormone Mechanism
Ngamkajornwiwat, P., Teerakittikul, P., & Manoonpong, P. (2018) · Lecture notes in computer science · doi:10.1007/978-3-319-97628-0_18 · Cited by 7
Deep Bayesian Optimization on Attributed Graphs
Cui, J., Yang, B. X., & Hu, X. (2019) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v33i01.33011377 · Cited by 8
When Your Robot Breaks: Active Learning During Plant Failure
Schrum, M., & Gombolay, M. (2019) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2019.2961598 · Cited by 8
“Prior work has sought to address safely learning a damage model [2], [3].”
BackgroundMethodInfluentialNakashima, S., Shirai, T., Kawaharazuka, K., Asano, Y., Kakiuchi, Y., Okada, K., & Inaba, M. (2019) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros40897.2019.8967949 · Cited by 6
“For example, A robot that learns optimal motion pattern after fracture was developed [5].”
BackgroundLearning from Sparse and Delayed Rewards with a Multilayer Spiking Neural Network
Chevtchenko, S. F., & Ludermir, T. B. (2020) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn48605.2020.9206846 · Cited by 6
“and MAP-Elites [32] are two promising methods for learning in complex action spaces and could be combined with the proposed spiking network.”
BackgroundMethodRobust Policy Search for Robot Navigation
Garcia-Barcos, J., & Martínez-Cantín, R. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2021.3060408 · Cited by 6
“The surrogate model can also be exploited for other purposes that can be useful in robotics or reinforcement learning scenarios, such as, guaranteeing a minimum outcome [19], detect and remove outliers [20] or incorporate prior information [17].”
BackgroundKatona, A., Franks, D. W., & Walker, J. A. (2021) · Conference on Artificial Life (ALIFE) · doi:10.1162/isal_a_00414 · Cited by 7
T-DominO: Exploring Multiple Criteria with Quality-Diversity and the Tournament Dominance Objective
Gaier, A., Stoddart, J., Villaggi, L., & Bentley, P. J. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-14721-0_19 · Cited by 8
“MAP-Elites [7,33], the most widely used QD algorithm, divides the feature space and searches for the best solution in each partition.”
Fast Design Space Exploration of Nonlinear Systems: Part I
Narain, S., Mak, E., Chee, D., Englot, B., POCHIRAJU, K. V., Jha, N. K., & Narayan, K. (2021) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2021.3118963 · Cited by 10
“[46] for adapting robot gait to failures in the field.”
BackgroundMethodInfluential
Metadata
- Authors
- Cully, A., Clune, J., Tarapore, D., & Mouret, J.-B.
- Affiliations
- Centre National de la Recherche Scientifique (FR)
- University of Wyoming (US)
- Corresponding author
- Mouret, Jean-Baptiste
- Venue
- Nature
- Published
- 2015-05-27
- Volume
- 521
- Issue
- 7553
- Pages
- 503-507
- ISSN
- 0028-0836 (ISSN-L)
- 1476-4687
- OpenAlex
- W1738827650
- Cited by
- 979
- References
- 40
- Citations by year
- 2026: 51 · 2025: 76 · 2024: 88 · 2023: 90 · 2022: 100 · 2021: 115 · 2020: 129 · 2019: 112 · 2018: 78 · 2017: 78 · 2016: 48 · 2015: 13
- FWCI
- 70.47
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- www.springer.com/tdm (tdm)
- www.springer.com/tdm (vor)
- Locations
- Nature (publishedVersion)
- PubMed (publishedVersion)
- ePrints Soton (University of Southampton) (acceptedVersion)
- arXiv (Cornell University) (submittedVersion)
- HAL (submittedVersion, other-oa)
- RePEc: Research Papers in Economics (submittedVersion)
- arXiv (Cornell University)
- Funders
- Agence Nationale de la Recherche (ANR-12-JS03-0009)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Reinforcement Learning in Robotics
- Physical Sciences › Engineering › Mechanical Engineering › Modular Robots and Swarm Intelligence
- Physical Sciences › Computer Science › Artificial Intelligence › Evolutionary Algorithms and Applications
- Keywords
- compensatory behavior, trial-and-error learning, autonomous mobile robots, legged robots, robot manipulators, prior knowledge, robot robustness, search and rescue robots, disaster response
- MeSH terms
- Adaptation, Physiological, Algorithms, Animals, Artificial Intelligence, Behavior, Animal, Dogs, Extremities, Motor Skills, Robotics, Time Factors, Biomimetics
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.